Supersonic panel flutter is a self-excited vibration, where energy is transferred into thin flexible panels, such as skin panels in supersonic vehicles, from the (supersonic) flow around them, which can lead to their rapid fatigue failure. Relevant examples of this include the failure of skin panels in the X-15 vehicle, or violent oscillations in non-adaptable nozzles during start-up, such as the space shuttle main booster. With the contemporary trends of decreasing weight and increasing operational speeds in the design of next generation aircraft and launch vehicles, such structures will become more common, and flutter needs to be inherently accounted for in the design.
In recent years, advancements in numerical techniques, such as Computational Fluid Dynamics (CFD) and the Finite Element Method (FEM), have led to numerous studies that address the effect of non-linearities, such as flutter behaviour in the transonic and hypersonic regime. In spite of that, experimental data needed to validate these studies is limited, based on coarse pointwise measurement techniques, and only focused on the structural response of the panel rather than on the entire Fluid Structure Interaction (FSI). The use of simultaneous, full-field and non-intrusive measurements can be used to address these issues. To this extent, panel designs were developed and an experimental campaign was proposed where Digital Image Correlation (DIC) and Schlieren are simultaneously used by means of high speed cameras to study classical panel flutter at Mach 2 inside the ST-15 wind tunnel at the Delft University of Technology. Additionally, a Laser Doppler Vibrometer (LDV) was used to validate the DIC measurements. The experiment allowed for a comparison with well understood theory, such that the setup can be proven to be robust, and can therefore later be used to assess the effect of panel flutter non-linearities or other supersonic FSIs.
Synchronisation between DIC and schlieren was demonstrated through a matching frequency spectrum, and maximum correlation inside corresponding pixel windows between panel displacement and schlieren grey-scale fluctuations with a zero time-lag. One finding was that during the panel down-stroke phase, a travelling wave character was observed which has previously been found only under transonic conditions. Furthermore, at moderate dynamic flutter pressures, a typical second bending panel flutter mode shape was observed, which grew into a single bending mode again with increasing higher dynamic flutter pressures. From the spectral analysis it was found that all configurations fluttered at the same frequency of 770 [Hz], independent of dynamic flutter pressure or edge conditions. This was higher than expected from analysis. These combined observations led to the conclusion that, in the current facility, a frequency lock-in occurs due to merging of the flutter phenomenon with a wind tunnel resonance vibration.
These observations provide an insight into combining these experimental techniques to study supersonic FSI behaviour, and particularly into certain unexpected behaviour encountered at the TU Delft ST-15 wind tunnel. This can be used to further tune future experiments and guide researchers to obtain a better understanding in the non-linearities found in panel flutter.